diff --git a/tensorboard/plugins/hparams/backend_context.py b/tensorboard/plugins/hparams/backend_context.py
index 806e9b236d6..98959498d5e 100644
--- a/tensorboard/plugins/hparams/backend_context.py
+++ b/tensorboard/plugins/hparams/backend_context.py
@@ -403,7 +403,7 @@ def compute_metric_infos_from_data_provider_session_groups(
         self, ctx, experiment_id, session_groups
     ):
         session_runs = set(
-            f"{s.experiment_id}/{s.run}"
+            f"{s.experiment_id}/{s.run}" if s.run else s.experiment_id
             for sg in session_groups
             for s in sg.sessions
         )
diff --git a/tensorboard/plugins/hparams/backend_context_test.py b/tensorboard/plugins/hparams/backend_context_test.py
index 0273eafe642..e11bfed7a6e 100644
--- a/tensorboard/plugins/hparams/backend_context_test.py
+++ b/tensorboard/plugins/hparams/backend_context_test.py
@@ -561,7 +561,7 @@ def test_experiment_from_data_provider_discrete_string_hparam(self):
         """
         self.assertProtoEquals(expected_exp, actual_exp)
 
-    def test_experiment_from_data_provider_session_group(self):
+    def test_experiment_from_data_provider_session_groups(self):
         self._mock_tb_context.data_provider.list_tensors.side_effect = None
         # The sessions chosen here mimic those returned in the implementation
         # of _mock_list_tensors. These work nicely with the scalars returned
@@ -614,6 +614,44 @@ def test_experiment_from_data_provider_session_group(self):
         """
         self.assertProtoEquals(expected_exp, actual_exp)
 
+    def test_experiment_from_data_provider_session_group_without_run_name(self):
+        self._mock_tb_context.data_provider.list_tensors.side_effect = None
+        self._hyperparameters = provider.ListHyperparametersResult(
+            hyperparameters=[],
+            session_groups=[
+                provider.HyperparameterSessionGroup(
+                    root=provider.HyperparameterSessionRun(
+                        experiment_id="exp/session_1", run=""
+                    ),
+                    # The entire path to the run is encoded in the experiment_id
+                    # to allow us to test empty run name while still generating
+                    # metric_infos.
+                    sessions=[
+                        provider.HyperparameterSessionRun(
+                            experiment_id="exp/session_1", run=""
+                        ),
+                    ],
+                    hyperparameter_values=[],
+                ),
+            ],
+        )
+        actual_exp = self._experiment_from_metadata()
+        expected_exp = """
+            metric_infos: {
+              name: {group: '', tag: 'accuracy'}
+            }
+            metric_infos: {
+              name: {group: '', tag: 'loss'}
+            }
+            metric_infos: {
+              name: {group: 'eval', tag: 'loss'}
+            }
+            metric_infos: {
+              name: {group: 'train', tag: 'loss'}
+            }
+        """
+        self.assertProtoEquals(expected_exp, actual_exp)
+
     def test_experiment_from_data_provider_old_response_type(self):
         self._hyperparameters = [
             provider.Hyperparameter(